Investigating the phenology of rubber trees (Hevea brasiliensis), a widely cultivated crop in tropical regions exhibiting distinct phenological patterns, is crucial for optimizing plantation management, enhancing production efficiency, and understanding the implications of global climatic change on tropical vegetation dynamics. The purpose of this study is to explores the spatiotemporal variation characteristics and driving factors of rubber phenology over the past two decades on Hainan Island, China’s second-largest rubber plantation base. We applied by integrating multi-source remote sensing imagery processed through the Google Earth Engine (GEE) cloud platform with various analytical methods including the seasonal amplitude method, Savitzky-Golay (S-G) filtering technique, partial correlation analysis, Sen’s slope, and Mann-Kendal test. The results showed that rubber phenology exhibited significant interannual trends and spatial heterogeneity. Specifically, the start of the growing season (SOS) mainly occurred from early to late March (day of year, DOY 60-81), with a trend of advancing by 1.1 days per decade. The end of the growing season (EOS) mostly occurred from late January to early February (DOY 392-406 counted from the previous year), delaying by 3.7 days per decade. The length of the growing season (LOS) lasted 10 to 11 months. Precipitation and topography significantly influenced rubber phenology, with SOS advancing as slope increased and elevation decreased, and increased precipitation accelerating both SOS and EOS at a rate of 1.00 day per 100 mm. Pre-seasonal climatic factors, particularly temperature in February and precipitation in February and March for SOS, and precipitation in January and temperature in February for EOS, were closely related to phenological changes. These findings elucidate the spatiotemporal patterns and underlying drivers of rubber tree phenological changes, offering valuable insights for optimizing rubber plantation management and informing tropical vegetation conservation efforts. Future research should prioritize investigating the impact of global climate change on rubber phenology and developing adaptive strategies to ensure sustainable development in tropical regions.

To date, few studies have assessed the impact of forest conversion or seasonal changes on soil microbial community assembly. To fill this research gap, 16S rRNA and ITS gene sequences were used to evaluate the effects of forest conversion and seasonal changes on the assembly of bacterial and fungal communities using 260 soil samples collected from tropical rainforest and rubber plantation sites across Hainan Island, South China. A majority (~60%) of observed OTUs conformed with neutral model expectations, indicating that neutral processes were important for the assembly of soil microbial communities. For bacterial communities, the NST (normalized stochasticity ratio) was higher in the tropical rainforest (0.746 in the dry season, 0.684 in the rainy season) versus rubber plantation sites (0.647, 0.584), regardless of season. Thus, forest conversion decreased the importance of stochasticity for soil bacterial community assembly. For fungal communities, rubber plantation communities showed greater stochasticity (NST = 0.578) than rainforest communities (NST = 0.388) in the dry season, but the reverse was true in the rainy season (NST = 0.852 for rubber plantations; NST = 0.978 for rainforest). Both the NST results and structural equation modeling showed that bacterial communities were more stochastic in the dry season, while fungal communities were more stochastic in the rainy season; the effects of seasonal changes on assembly therefore differed between bacterial and fungal communities. More importantly, forest conversion did not have a direct impact on the assembly of bacterial or fungal communities, but exerted indirect effects via soil pH and soil AK.

The effects of forest conversion from natural forest to agricultural system on soil microbial composition still need further study. Especially, impact on soil function after forest conversion is not yet known. In this study, by using metagenomic sequencing as well as 16S and ITS sequencing technology, we evaluated the soil microbial composition, diversity and functions based on a large number of soil samples of tropical rainforest and rubber plantation across the whole island of Hainan, south China. The results showed that (1) forest conversion changed microbial composition from bacterial groups of Proteobacteria to Chloroflexi, and fungal groups from Basidiomycota to Ascomycota. (2) The bacterial alpha diversity, beta diversity as well as the total diversity did not decrease after forest conversion. However, beta diversity of fungal community reduced resulting a net loss of total OTU richness. (3) There was no difference in soil functional compositions and diversity between rubber plantations and rainforest, however, the relative gene abundance of most COG functions, KEGG functions, CAZy functions as well as Antibiotic gene were significantly different between rubber plantation and tropical rainforest. (4) Soil pH and environmental heterogeneity were the main driver for microbial taxonomic composition and gene functional composition. Land use did not result in changes of functional gene composition, but the relative abundance of functional gene. The changed relative abundance gene would alter the ecosystem processes. In conclusion, our results confirmed that land use changes alter the soil microbial community structure and can have profound effects on ecosystem functions and processes.